Freezemart - E-commerce Frozen Food Website with Recommendation System
Project Overview
Freezemart is a frozen food e-commerce platform featuring a content-based recommendation system powered by TF-IDF and Cosine Similarity, and integrated with Xendit payment gateway for seamless in-app transactions.
My Role
Full-Stack Developer responsible for implementing the frontend UI, backend logic, integrating the recommendation algorithm, and configuring Xendit payment gateway to enable secure online payments.
Challenges & Solutions
Designing and integrating a performant recommendation system, ensuring real-time product suggestions without impacting page load times, maintaining consistency across the Laravel and Flask components, and implementing a secure payment flow via Xendit.
Project Development Process
User Research & Data Collection
Gathered user preferences and historical purchase data, cleaned and preprocessed datasets for the recommendation algorithm.
System Architecture & Planning
Outlined the full-stack architecture, defined data flow between Laravel frontend, Flask recommendation API, and Xendit payment service, and planned database schemas.
Recommendation Engine Development
Implemented TF-IDF vectorization and Cosine Similarity matching in Python, deployed as a Flask microservice for product suggestions.
Payment Integration & Frontend Development
Integrated Xendit payment gateway into the Laravel frontend, handled payment callbacks, and built responsive UI components in Blade and TailwindCSS for checkout flow.
Testing & Optimization
Conducted unit and integration tests for recommendation accuracy and payment flow, optimized query performance and caching strategies to minimize latency.
Deployment & Maintenance
Deployed application on a Linux server, configured CI/CD pipelines, monitored system performance and payment logs, and iterated based on user feedback.